Nucleic acid-based therapeutics have the potential to provide potent and highly specific treatments for a variety of human ailments including cancer, infectious diseases, and autoimmune disorders. However, rapid nuclease degradation and clearance from the bloodstream precludes systemic application of unmodified nucleic acids. One approach to overcome this problem is to encapsulate them inside a liposomal or polymeric carrier that can transport the nucleic acid payload safely through the bloodstream and deliver it to a particular cell of interest, such as cancer cells in a tumor metastasis. The delivery vehicle must be engineered to have the following characteristics: (i) be small enough to extravasate and exhibit adequate tissue penetration, yet still avoid rapid renal clearance; (ii) protect the nucleic acid from degradation, yet willingly release it upon arrival at the proper site; and (iii) avoid nonspecific interactions and opsonization while still providing specific targeting to a given cell. We are developing a cyclodextrin-based polycation (CDP) that can interact with nucleic acid molecules to form polyplexes that achieve a favorable balance in these design parameters. Here, we use the CDP delivery vehicle as a model system to highlight several key criteria for the design of targeted nucleic acid delivery vehicles in general. First, the delivery vehicle must be small enough to allow transport and diffusion within tissues without rapid renal clearance, while still being large enough to carry a substantial nucleic acid payload. It is believed that nanoparticles around 100 nm possess an optimum balance with regard to transport. Using dynamic and multi-angle light scattering, we show that polyplexes formed with CDP and either plasmid DNA or small interfering RNA (siRNA) have effective diameters of 50-100 nm and molecular weights of 10^6-10^7 g/mol. Calculations show that a polyplex with a molecular weight of 10^7 g/mol contains approximately 100-200 siRNA molecules. This ability of the targeted delivery vehicles to deliver hundreds of siRNA molecules into a cell with each uptake event presents a distinct advantage over the direct conjugation of targeting ligands to individual siRNA molecules. Second, the delivery vehicle must protect its nucleic acid payload from nuclease degradation in the presence of physiological fluids. Our results demonstrate that although nuclease-stabilized siRNAs show enhanced serum stability, they have no significant effect on the duration or magnitude of gene knockdown once inside cells. The ability of CDP to protect its nucleic acid payload explains why we can achieve knockdown of target genes in mice after systemic delivery of targeted polyplexes containing unmodified siRNA. Therefore, delivery vehicles that can protect the nucleic acid from degradation obviate the need for expensive or difficult chemical modifications to stabilize nucleic acid molecules even for systemic application. Third, delivery vehicles must display a delicate balance to prevent nonspecific uptake while maximizing uptake by the tumor cells. Surface decoration with hydrophilic molecules such as polyethylene glycol (PEG) can reduce nonspecific interactions, as revealed by our observations that PEGylated polyplexes do not aggregate in physiological salt solutions, do not cause any observable erythrocyte aggregation, and do not exhibit complement activation at the concentrations required for in vivo delivery of siRNA. We recently demonstrated the ability of transferrin-targeted polyplexes formed with CDP and siRNA against the oncogenic fusion gene EWS-FLI1 to inhibit tumor growth in a murine model of Ewing's sarcoma. Polyplexes formulated identically except without the transferrin targeting ligand did not inhibit tumor growth, highlighting the importance of targeting ligands for specific uptake by tumor cells. One perhaps under-utilized property of targeted delivery vehicles is the multivalent nature of their target binding that can lead to increased affinity through avidity effects. In competitive uptake experiments using flow cytometry, we show that transferrin-targeted polyplexes exhibit an enhanced affinity for the transferrin receptor compared to free transferrin conjugates alone. The affinity can be tuned by modulating the targeting ligand density, a feature that can be particularly important for uptake by and penetration within tumors since high-affinity ligands can lead to a so-called binding-site barrier. Imaging techniques such as MRI, PET, and confocal microscopy will be used to examine these effects after systemic delivery of targeted polyplexes in murine cancer models. These results underscore the importance of a rational approach in delivery vehicle design and emphasize the need for continued studies to further understand the role of multivalent binding effects on the uptake and distribution of targeted delivery vehicles.